Economics

I study repeated communication games between a patient sender and a sequence of receivers. The sender has persistent private information about his psychological cost of lying, and in every period, can privately observe the realization of an i.i.d. state before communication takes place. I characterize every type of sender's highest equilibrium payoff. When the highest lying cost in the support of the receivers' prior belief approaches the sender's benefit from lying, every type's highest equilibrium payoff in the repeated communication game converges to his equilibrium payoff in a one-shot Bayesian persuasion game. I also show that in every sender-optimal equilibrium, no type of sender mixes between telling the truth and lying at every history. When there exist ethical types whose lying costs outweigh their benefits, I provide necessary and sufficient conditions for all non-ethical type senders to attain their optimal commitment payoffs. I identify an outside option effect through which the possibility of being ethical decreases every non-ethical type's payoff.

We study a repeated game with payoff externalities and observable actions where two players receive information over time about an underlying payoff-relevant state, and strategically coordinate their actions. Players learn about the true state from private signals, as well as the actions of others. They commonly learn the true state (Cripps et al., 2008), but do not coordinate in every equilibrium. We show that there exist stable equilibria in which players can overcome unfavorable signal realizations and eventually coordinate on the correct action, for any discount factor. For high discount factors, we show that in addition players can also achieve efficient payoffs.

I study a social learning model in which the object to learn is a strategic player's endogenous actions rather than an exogenous state. A patient seller faces a sequence of buyers and decides whether to build a reputation for supplying high quality products. Each buyer does not have access to the seller's complete records, but can observe all previous buyers' actions, and some informative private signal about the seller's actions. I examine how the buyers' private signals affect the speed of social learning and the seller's incentives to establish reputations. When each buyer privately observes a bounded subset of the seller's past actions, the speed of learning is strictly positive but can vanish to zero as the seller becomes patient. As a result, reputation building can lead to low payoff for the patient seller and low social welfare. When each buyer observes an unboundedly informative private signal about the seller's current-period action, the speed of learning is uniformly bounded from below and a patient seller can secure high returns from building reputations. My results shed light on the effectiveness of various policies in accelerating social learning and encouraging sellers to establish good reputations.

A patient player privately observes a persistent state that directly affects his myopic opponents' payoffs, and can be one of the several commitment types that plays the same mixed action in every period. I characterize the set of environments under which the patient player obtains at least his commitment payoff in all equilibria regardless of his stage-game payoff function. Due to interdependent values, the patient player cannot guarantee his mixed commitment payoff by imitating the mixed-strategy commitment type, and small perturbations to a pure commitment action can significantly reduce the patient player's guaranteed equilibrium payoff.

At the zero lower bound, the New Keynesian model predicts that output and inflation collapse to implausibly low levels, and that government spending and forward guidance have implausibly large effects. To resolve these anomalies, we introduce wealth into the utility function; the justification is that wealth is a marker of social status, and people value status. Since people partly save to accrue social status, the Euler equation is modified. As a result, when the marginal utility of wealth is sufficiently large, the dynamical system representing the zero-lower-bound equilibrium transforms from a saddle to a source---which resolves all the anomalies.

McFadden and Richter (1991) and later McFadden (2005) show that the Axiom of Revealed Stochastic Preference characterizes rationalizability of choice probabilities through random utility models on finite universal choice spaces. This note proves the result in one short, elementary paragraph and extends it to set valued choice. The latter requires a different axiom than is reported in McFadden (2005).

In a decision problem comprised of multiple choices, a person may fail to take into account the interdependencies between her choices. To understand how people make decisions in such problems we design a novel experiment and revealed preference tests that determine how each subject brackets her choices. In separate portfolio allocation under risk, social allocation, and induced-utility shopping experiments, we find that 40-43\% of our subjects are consistent with narrow bracketing while only 0-15\% are consistent with broad bracketing. Classifying subjects while adjusting for models' predictive precision, 73\% of subjects are best described by narrow bracketing, 14\% by broad bracketing, and 5\% by intermediate cases.

This note strengthens the main result of Lagziel and Lehrer (2019) (LL) "A bias in screening" using Chambers Healy (2011) (CH) "Reversals of signal-posterior monotonicity for any bounded prior". LL show that the conditional expectation of an unobserved variable of interest, given that a noisy signal of it exceeds a cutoff, may decrease in the cutoff. CH prove that the distribution of a variable conditional on a lower signal may first order stochastically dominate the distribution conditional on a higher signal. The nonmonotonicity result is also extended to the empirically relevant exponential and Pareto distributions, and to a wide range of signals.

We study the regulation of a monopolistic firm using a robust-design approach. We solve for the policy that minimizes the regulator's worst-case regret, where the regret is the difference between his complete-information payoff minus his realized payoff. When the regulator's payoff is consumers' surplus, it is optimal to impose a price cap. The optimal cap balances the benefit from more surplus for consumers and the loss from underproduction. When his payoff is consumers' surplus plus the firm's profit, he offers a piece-rate subsidy in order to mitigate underproduction, but caps the total subsidy so as not to incentivize severe overproduction.

We study sequential search without priors. Our interest lies in decision rules that are close to being optimal under each prior and after each history. We call these rules dynamically robust. The search literature employs optimal rules based on cutoff strategies that are not dynamically robust. We derive dynamically robust rules and show that their performance exceeds 1/2 of the optimum against binary environments and 1/4 of the optimum against all environments. This performance improves substantially with the outside option value, for instance, it exceeds 2/3 of the optimum if the outside option exceeds 1/6 of the highest possible alternative.